Rockbox General > Announcements

BUTTON CHANGES - REPORT PROBLEMS HERE

<< < (41/42) > >>

nls:
I have found a minor annoyance in the WPS context menu. If I press right on the reshuffle playlist item in the menu it just goes back to the WPS and does not reshuffle, works fine if I press NAVI though...

JdGordon:
nls: ok thats wierd.. I seem to be getting the opposite here (h300). righ works, navi doesnt..

actually.. no, i tihnk both work, but for some reason its more instant with right!

nls:
Yes you're right both works for me now but it takes a good 10 seconds for the wps nex song and the playlist viewer to update even with a playlist of about 15 tracks... guess it's not a button problem then!  ;)

roolku:
Following on from our discussion on IRC regarding button presses being eaten after exiting certain plugins I have done more experiments.

A plugin only containing:


--- Code: ---while (rb->button_get(false) != BUTTON_OFF);
--- End code ---
is okay


--- Code: ---    while (rb->button_get(false) != (BUTTON_OFF | BUTTON_REL));
--- End code ---
is not okay


--- Code: ---while (rb->button_get(false) != BUTTON_OFF);
rb->splash(HZ/3, true, "Bye now!");
--- End code ---
is not okay for a short press


--- Code: ---while (rb->button_get(false) != BUTTON_OFF);
rb->splash(HZ/3, true, "Bye now!");
--- End code ---
is okay if I hold the button long enough for the splash to disappear

So I am guessing that there needs to be a button code in the queue (i.e. the button release) when exiting the plugin that is then eaten by the action code. If there isn't a "victim" left, the action code will eat the next intentional press available.

I tried to 'repair' calculator.rock by removing the splash screen at the end and it 'fixed' it. However this is not the correct way of solving it, it needs to be fixed outside the plugin.

EDIT:

I found a call to button_clear_queue(); in plugin_load(); immediately after the plugin returns. Removing it fixed the problem for the vast majority of plugins. Some (including the calculator) still have the problem though. Also this call was probably there for a reason, so I don't know what side-effect its removal will have on other targets.



JdGordon:
fix committed

Navigation

[0] Message Index

[#] Next page

[*] Previous page

Go to full version